Re: (attn doc team) Re: Diatonic notation system

[Hans Aberg]

On 10 Dec 2008, at 15:04, Graham Percival wrote:

> You'll want to pay special attention to ly/arabic.ly, and make
> sure you understand everything that's happening in there.

Have you looked it into yourself? - It goes on like

bayati = #`(
    (0 . 0)
    (1 . ,SEMI-FLAT)
    (2 . ,FLAT)
    (3 . 0)
    (4 . 0)
    (5 . ,FLAT)
    (6 . ,FLAT)
  )

Now, everyone agrees that in Arab music, the intermediate pitch isn't  
that - in fact the guy who write it said he was taught to lower it,  
which agree with the values Farhat uses in Persian music (using koron  
and sori; see below).

In addition, Arab music may not even be played in a micro-toning of  
E12, but E53 or Pythagorean tuning. In that case for example F# and  
Gb no longer agree, since M = 9, m = 4, and this accidentals alter  
with M - m = 5 commas (or E53 tonesteps). For example C to C# is 5  
commas, and C to Db is 4 commas. In Persian music, the intermediate  
pitch is described by Farhat using a neutral second n, which in E53  
gets the value 6 commas. There result a koron p that lowers 3 commas,  
and a sori > that raises 2 commas. Note the asymmetry: the sum of the  
p and > intervals is 5 commas = M - m. This so in general.

There are two problems: does the E24 setting in transposing (or  
choosing a key) risk say E- (half-flat) be confused with D#+ (sharp  
and half-flat). - This I do not know, because it depends on how  
LilyPond computes it.

If there is no risk for such confusion, then E24 is just fine, except  
for that the MIDI files will not sound good.

So in order to lessen the risk for such confusion, I suggested using  
E36. The motivation comes from Persian music, where one suggestion is  
that Farhat's neutral second n has the value 27/25. If I keep that  
value, then E36 approximates it well, and it also sounds good to me  
when trying it in Scala.

Of course, that is all attempts to find a workaround in LilyPond,  
which requires explicit number values plugged in for intervals.

  Hans